Computer-Implemented Automated Authorization System Using Natural Language Processing

1. An interrogatory development system comprising: at least one processor; and a memory coupled to the at least one processor, wherein the memory stores: a question repository that includes a plurality of questions corresponding to a plurality of decision trees, wherein each of the plurality of questions corresponds to a set of the plurality of decision trees; a distance matrix that encodes a distance between each pair of questions in the plurality of questions; and instructions that, upon execution, cause the at least one processor to, in response to receiving a request for a new question: convert the new question into a set of tokens; for each question of the plurality of questions, determine a distance between the question and the new question by: for each token of the new question, determining a minimum distance between the token and tokens of the question; and summing the minimum distances to calculate the distance; perform cluster analysis on the distance matrix, wherein performing cluster analysis includes (i) normalizing the distance matrix and (ii) applying a hierarchical clustering process to the normalized distance matrix; in response to the cluster analysis indicating the new question clusters with at least one of the plurality of questions, generate an alternative question proposal and transmit the alternative question proposal as a response to the request; and in response to the cluster analysis indicating the new question clusters with zero of the plurality of questions, add the new question to the plurality of questions and transmitting a question added message as the response to the request.

2. The interrogatory development system of claim 1 wherein the alternative question proposal is generated based on a selected question of the at least one of the plurality of questions.

3. The interrogatory development system of claim 2 wherein the instructions, upon execution, cause the at least one processor to assign ranks to the at least one of the plurality of questions and choose a highest-rank one of the at least one of the plurality of questions as the selected question.

4. The interrogatory development system of claim 3 wherein: the ranks are based on empirical data gathered on the at least one of the plurality of questions and the empirical data indicates how frequently data can be automatically obtained for each of the at least one of the plurality of questions.

5. The interrogatory development system of claim 4 wherein the empirical data indicates how frequently automatically obtained data for each of the at least one of the plurality of questions must be modified.

6. The interrogatory development system of claim 1 wherein the instructions, upon execution, cause the at least one processor to incorporate the distances between the questions and the new question into the distance matrix.

7. The interrogatory development system of claim 1 wherein: determining a first minimum distance between a first token and tokens of the question includes determining a set of distances and selecting a smallest one of the set of distances as the first minimum distance and each distance of the set of distances indicates distance between a vector representation of the first token and a vector representation of a respective token of the question.

8. The interrogatory development system of claim 7 wherein a distance between the vector representation of the first token and a vector representation of a second token is calculated as a Euclidean distance between the vector representation of the first token and the vector representation of the second token.

9. The interrogatory development system of claim 1 wherein: the memory stores a vector data store including a plurality of vector representations and each vector representation of the plurality of vector representations corresponds to a respective token.

10. The interrogatory development system of claim 1 wherein converting the new question into a set of tokens includes: performing word stemming on words in the new question; removing punctuation from the new question; and selectively removing words from the new question, wherein each token of the set of tokens corresponds to a remaining word of the new question.

11. The interrogatory development system of claim 10 wherein: a word list specifies words to be removed from the new question and the word list includes medical terms.

12. The interrogatory development system of claim 11 wherein the medical terms includes brand names of prescription drugs and generic names of prescription drugs.

13. The interrogatory development system of claim 1 wherein the hierarchical clustering process includes k-means clustering.

14. A method comprising: maintaining a question repository that includes a plurality of questions corresponding to a plurality of decision trees, wherein each of the plurality of questions corresponds to a set of the plurality of decision trees; maintaining a distance matrix that encodes a distance between each pair of questions in the plurality of questions; and in response to receiving a request for a new question: converting the new question into a set of tokens; for each question of the plurality of questions, determining a distance between the question and the new question by: for each token of the new question, determining a minimum distance between the token and tokens of the question; and summing the minimum distances to calculate the distance; performing cluster analysis on the distance matrix, wherein performing cluster analysis includes (i) normalizing the distance matrix and (ii) applying a hierarchical clustering process to the normalized distance matrix; in response to the cluster analysis indicating the new question clusters with at least one of the plurality of questions, generating an alternative question proposal and transmitting the alternative question proposal as a response to the request; and in response to the cluster analysis indicating the new question clusters with zero of the plurality of questions, adding the new question to the plurality of questions and transmitting a question added message as the response to the request.

15. The method of claim 14 wherein the alternative question proposal is generated based on a selected question of the at least one of the plurality of questions.

16. The method of claim 15 further comprising: assigning ranks to the at least one of the plurality of questions and choosing a highest-rank one of the at least one of the plurality of questions as the selected question.

17. The method of claim 16 wherein: the ranks are based on empirical data gathered on the at least one of the plurality of questions; the empirical data indicates how frequently data can be automatically obtained for each of the at least one of the plurality of questions; and the empirical data indicates how frequently automatically obtained data for each of the at least one of the plurality of questions must be modified.

18. The method of claim 14 further comprising incorporating the distances between the questions and the new question into the distance matrix.

19. The method of claim 14 wherein: determining a first minimum distance between a first token and tokens of the question includes determining a set of distances and selecting a smallest one of the set of distances as the first minimum distance and each distance of the set of distances indicates Euclidean distance between a vector representation of the first token and a vector representation of a respective token of the question.

20. The method of claim 14 wherein converting the new question into a set of tokens includes: performing word stemming on words in the new question; removing punctuation from the new question; and selectively removing words from the new question, wherein each token of the set of tokens corresponds to a remaining word of the new question.